Closed fsteimke closed 5 months ago
The whitespace looks fine, thanks.
I'm confused about the note though. The 10-xinclude.xsl
transform preprocessing step is supposed to expand XIncludes. It does so with the sinclude extension function. Can you describe the circumstance when you needed to use the Xerces implementation (which is what the xi:on
option enables).
If your XIncludes are simple, using the parser flag might be marginally more efficient, but it supports such a narrow range of fragment identifiers that I've mostly abandoned it.
That is a misunderstanding. It was never about Xerces. I was convinced that no xinclude processing takes place without the -xi
switch. As described in Issue #430, i had noticed with a certain stylesheet that xinclude did not take place without this switch. I had assumed that this was always the case.
However, I have now checked this on a very simple case and found that xinclude was correct processed without the switch. I have to investigate. Either I was simply mistaken, or there was something special about the customized stylesheet.
In any case, I withdraw this pull request because I was working on the wrong assumptions.
Adds a note in the reference guide for processing modular DocBook files.
I hope that there are no problems with whitespace which makes comparison difficult. Otherwise give me a hint.
Greetings, Frank